4 experiments were conducted concerning the development of visual accommodation in 1- to 3-month-old infants. In experiments 1 and 2 dynamic retinoscopy was used to measure accomodation responses at 3 stimulus distances. The results of experiment 1 revealed better accommodative capability from 1 to 3 months than reported originally. The procedure of experiment 2 was somewhat different but the results confirmed those of experiment 1. In experiment 3, accommodative responses at 7 stimulus distances were carefully measured in a small number of infants. These data provided estimates of the shape of infants' accommodation functions. In experiment 4, we used infrared photography to measure infants' pupil diameters while they viewed the stimuli of experiments 1 and 2. 2 simple hypotheses of the developmental mechanisms which underlie early accommodative development were considered. First, development of the motor component of the accommodative system might determine accommodative development. Second, development of the sensory component of the accommodative system might determine the observed development. The first hypothesis was tentatively rejected because it is inconsistent with some clinical findings. Evaluation of the second hypothesis involved calculating infants' depth of focus. We used those depth-of-focus values to predict how well infants of different ages should accommodate if their only limitation were in the sensory component of the accommodative system. The agreement between those predictions and observed accommodation was excellent, suggesting that changes in depth of focus in the first 3 months are largely responsible for growth in accommodation. The theoretical implications of this finding are discussed.